github-actions[bot] commented on code in PR #68125:
URL: https://github.com/apache/doris/pull/68125#discussion_r4068845794
##########
be/src/service/point_query_executor.cpp:
##########
@@ -633,6 +663,9 @@ Status PointQueryExecutor::_lookup_row_data() {
read_column, slot, row_ids, column,
storage_read_options, iter));
}
}
+
replace_point_query_read_time_hidden_columns(_reusable->read_time_hidden_columns(),
Review Comment:
This replacement is ineffective for a full-row-store row after ordinary
compaction. With a full row store, `get_missing_and_include_cids()` clears
`missing_cids`, so the requested hidden column is decoded only from JSONB.
Direct writes serialized VERSION/COMMIT_TSO there while they were still zero
placeholders, and `TYPE_COMPACTION` copies that row-store column without
`RowStoreFill`; the compacted output has a range such as `[2,3]`, so
`get_read_time_hidden_column_value()` returns `nullopt` here and leaves zero
unchanged. The same failure reaches the full-row-store lazy row-ID path. Please
read these hidden columns from their standalone columns (then apply singleton
replacement), or rebuild the JSONB during compaction, and add an exact-value
test after compacting at least two rowsets.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
